The astronomer would have survived a 12-hour deprivation of access to the page without problems, but in order to restore access, the ill-fated video had to be deleted
A photo: REUTERS
Another confirmation of the imperfection of artificial intelligence systems for moderating social media content: a six-second video with falling meteorites was recognized by the Twitter verification system as “too frank” and deprived the woman astronomer who published it of access to the page.
Mary McIntyre from Oxford, UK, complained that the meteor shower video she posted was hidden from users, and she herself was banned for “explicit content.”
The astronomer would have survived a 12-hour deprivation of access to the page without problems, but in order to restore access, the ill-fated video had to be deleted, which means an automatic admission of guilt, writes The Guardian.
Most of all, McIntyre was afraid that as a result of such actions, her leadership could receive a message about her “illegal” acts on the Web. She, as a person who periodically works with children, did not need such a spot on her reputation at all.
“It’s just some kind of madness … I don’t smile at all to have a record in my personal file that I distributed pornography – when in fact I didn’t do anything like that,” the astronomer said.
In the end, after the intervention of the media, the truth prevailed, and McIntyre again gained access to her page. It took her almost three months.
It is noteworthy that the astronomer’s colleagues, who published the same video at her request, had no problems. Apparently, by that time, artificial intelligence considered itself satisfied.